Structure-preserving model reduction of second-order systems by Krylov subspace methods

In this paper, structure-preserving model reduction methods for second-order systems are investigated. By introducing an appropriate parameter, the second-order system is represented by a strictly dissipative realization and the H₂ norm of the strictly dissipative system is discussed. Then, based on the Krylov subspace techniques, two model reduction methods are proposed to reduce the order of the strictly dissipative system. Further, the reduced second-order systems are obtained. Moreover, according to the factorization of the error system, the H₂ error bounds are represented by the Kronecker product and the vectorization operator. Finally, two numerical examples illustrate the efficiency of our methods.